The headline is scary: "Maternal exposure to anti-depressant SSRIs linked to autism in children." The Washington Post asserted that a study published Monday provides the "strongest evidence yet" that antidepressants during pregnancy may be linked with autism spectrum disorders in children. A press release was less nuanced: "Taking antidepressants during pregnancy increases risk of autism by 87 percent."
In reality, it's complicated. The study published Monday adds to a large literature showing that infants born to moms with health problems face higher risks than infants born to moms who are well. In this particular study, moms who had a mental health condition that was severe enough to require treatment with an antidepressant were more likely than women who did not require medication to have a child with a billing code for a developmental disorder.
Importantly, the study defined "autism" as a medical service claim or hospital code for autism or several related diagnoses. When the authors looked only at cases of autism that were confirmed by psychiatrists, link with antidepressants was no longer statistically significant. That means the study found that taking medications for depression is linked with having a child evaluated for autism, but not with actually having autism.
But that's not the fundamental limitation of study. The key problem is that women who take a medication when they are pregnant have a reason for taking it. Blaming the outcome on the medication, without considering the underlying disease, is like saying that umbrellas cause flooding. Taking away the umbrellas -- "not treating the rain" -- does not prevent flooding; it just means that people get soaked.
Furthermore, the authors did not compare women taking antidepressants with women with similarly severe depression who stopped their medication. When I am talking with my patients, that is the key question -- if I have a patient with a longstanding history of depression that is being treated with medication, is it better for her and her baby to continue treatment, or to risk worsening depression by stopping her medication?
The authors tried to sort this out by comparing women with a history of depression with women who were taking antidepressants. However, these two groups are very different. A woman with a "history of depression" might have been depressed in high school when her grandparent died, and then been fine thereafter. She is completely different from a woman who is continuing to take antidepressants during pregnancy because she has severe major depressive disorder and a history of a suicide attempt.
There's evidence from well-designed studies that when women taking antidepressants are compared with women who share the same risk profile, medication is no longer linked with infant health problems. Moms with health problems have babies with health problems.
We also know that untreated depression is bad for babies -- infants born to depressed mothers have been shown to have changes in their neurotransmitters and behavior -- so it's entirely possible that if the mothers with severe depression stopped their treatment, rather than continuing it, their infants would be at higher risk than women who stayed on their medication.
Finally, the "87 percent increase" is wildly misleading. The risk of having a child evaluated for autism increased from 7 cases per 1,000 women not taking antidepressants to 12 cases per 1,000 women taking antidepressants. So if taking antidepressants actually caused the difference in risk (and that's far from proven by this study), the chance of a child having autism went up by 0.5 percent.
The take-away from this study is that we need to tailor treatment to the needs of each individual woman. As with any health condition in pregnancy, if non-pharmcologic treatments are effective, we should use them. We need to make sure that high-quality, evidence-based therapy is widely available. But for a woman who has severe disease, we need to ensure that she is adequately treated with a medication that controls her symptoms and allows her to function.
This week, I met a mom who said she didn't dare read the news about this study, because she was terrified. She had taken antidepressants during her second pregnancy, and that morning, she had asked her husband, imploringly, whether their son "looked like he has autism." We talked about the 0.5 percent absolute difference in risk, and the fact that untreated depression is bad for moms and babies. We talked about the fact that anxiety and depression, like diabetes and hypertension, are chronic health problems that require effective treatment for the wellbeing of mothers and babies.
My biggest concern about this study is that mothers will stop taking their medication to avoid the headlined "87 percent increased risk of autism" without the guidance and support of clinicians who can put the research in context. Although there may be risks of taking medication, there are substantial risks of stopping. In one study that followed mothers with severe depression, 68 percent of those who stopped their medication during pregnancy relapsed; of those who continued treatment, only 26 percent relapsed. To put this in context: If 1,000 women with severe depression stop taking their medications, 420 would have a relapse of major depression, and (if the current study is valid), 5 would avoid having a child evaluated for autism.
Depression and anxiety make pregnancy and parenthood more complex, whether or not mothers are taking an antidepressant. We need to ensure that every woman has access to skilled clinicians who understand the real risks and benefits of medication, therapy, and other treatment strategies, so that we can improve health and wellbeing for two generations.
Alison Stuebe, M.D., MSc, is a board certified Maternal-Fetal Medicine subspecialist at the University of North Carolina School of Medicine and Distinguished Scholar of Infant and Young Child Feeding at the Gillings School of Global Public Health. You can follow her on Twitter @astuebe.